Advanced Image Segmentation algorithm to analyze and detect Dermatitis Disease
issue 1

Advanced Image Segmentation algorithm to analyze and detect Dermatitis Disease

Prafulla N. Aerkewar 1, Dr. G. H. Agrawal 2,

1 Ph.D.Research scholar, B.D.College of Engg., Sevagram, Dist. Wardha,RTMNU,India

2 Professor, Department of EE, K.D.K.College of  Engineering,RTMNU, Nagpur,India.

1praful.yerkewar@gmail.com

2 ghagrawal66@yahoo.com

ABSTRACT:-The aim to publish this research paper to classify dermatitis disease using  an advanced  technique of image segmentation called k-means clustering method. The process of clustering image segmentation method extract different features of input test image of dermatitis disease and compare with database images features values. This method suggest appropriate procedure such that the all dermatitis disease having skin lesion on body are classified in to four category using k-means image segmentation and nntool of Matlab. Using the image segmentation technique and nntoo,  the analysis  and study of  the segmentation properties of skin lesions occurs in dermatitis diseas is possiblee. A skin lesion is a superficial growth or patch of the skin that does not resemble the area surrounding it. It have also been  proposed that which are suitable for the processing of various images for different types of patches for various skin diseases. The skin lesion in different dermatitis diseases are different in appearance and have different properties though  they looks similar in some circumstances. The main objective to classify the lesions of different dermatitis diseases based on its twelve parameters like contrast, Energy, Homogeneity etc where it would be able to classify the similar patch in to different disease.

Keywords:-Segmentation, Disease, Dermatitis  Lesion, Patch,, Leprosy, Vitiligo, Psoriasis, Feature Extraction, Feature Selection, Entropy, Autocorrelation,

I. INTRODUCTION

In day to day life, any person, rich or poor may cause patches on body due to many diseases such as Leprosy, Scar tissue, Contact dermatitis, Vitiligo, Psoriasis, Ring Worm, Eczema which may have  the main issue that the skin is affected and may have   the similar appearing symptoms like   patches on some part of body or all over the body. As the  patches on the body are looks similar but may cause by different reason. So due to similar appearance it may cause misdiagnosis and wrong treatment may be given.  Most of the time, it is complicated when such a thing happens complication like reaction of a treatment may affect a life of victim.

So to avoid such misdiagnosis, image segmentation using k-means suggest  the methodology for perfect diagnosis of the dermatological disease and helps in the analysis of such a disease, which can analyze the prior situation also. There are different image segmentation procedure that can be apply to build up such a powerful electronics tool which may create revolution in field of medical as well as engineering field. This effective technique work on analyzing the different parameter of skin lesion image, nntool and fuzzy logic of pervious database particular skin disease patch image.

II. METHODOLOGY

The database of Images is collected via personal contacts with a patients or finding person  having patch or patches on body. Also visited private skin care clinic and taken guidance regarding dermatitis disease. Total 70 people are contacted via visiting Maharogi seva Samiti, Dattapur  and some private clinic. Also personally contacted some people.    14 samples of each category have been taken for analysis. Out of 70 people 30 people does have any patch on body so they come under other category.  Out of remaining 40 people 30 found the patient of leprosy. Among 30 patient 14 sample are tested using algorithm and detected Leprosy correctly. Some Image samples are taken from patches image taken from internet. Sample images of Vitiligo and psoriasis also detected correctly. For analysis .  Above database is segmentated using k-Means clustering method. This method uses the  means clustering to cluster the dermatological lesions. By using the cluster values we classify the types of dermatological ulcers. Fuzzy K-Nearest neighbor classification is used to classify the types of dermatological ulcers. Each sample should be classified similarly to its surrounding samples, therefore, a unknown sample could be predicated by considering the  classification of its nearest neighbor samples. Given a sample set  values, a fuzzy M class partition of these vectors specify the membership degrees of each sample  corresponding to each class.

Figure 1:- Process Steps

Feature Extraction and Feature Selection:-

The collected database of skin lesions  are process using k-means clustering and median filter   and denoise image is generated. The denoised image forwarede for feature extraction and many feature like energy, Entropy, Autoco-orelation, homogeneity are extracted from database. The same 12  features are extracted from the test image to be diagnosis. Following are the  parameter are use for train and test images for feature extraction using below procedure.

Contrast:

Contrast is a measure of the local variations present in an image. If there is a large amount of variation in an image the P[i,j]’s will be concentrated away from the main diagonal and contrast will be high (typically k=2, n=1).                

Homogeneity:-

A homogeneous image will result in a co-occurrence matrix with a combination of high and low P[i,j]’s.

Where the range of gray levels is small the P[i,j] will tend to be clustered around the main diagonal.

Entropy:

Entropy is a measure of information content. It measures the randomness of intensity distribution.Such a matrix corresponds to an image in which there are no preferred gray level pairs for the distance vector d. Entropy is highest when all entries in P[i,j] are of similar magnitude, and small when the entries in P[i,j] are unequal.

Correlation:

Correlation is a measure of image linearity.

Correlation will be high if an image contains a considerable amount of linear structure.

Energy:

One approach to generating texture features is to use local kernels to detect various types of texture. After the convolution with the specified kernel, the texture energy measure (TEM) is computed by summing the absolute values in a local neighborhood: If n kernels are applied, the result is an n-dimensional feature vector at each pixel of the image being analyzed

Maximum Probability:

This is simply the largest entry in the matrix, and corresponds to the strongest response. This could be the maximum in any of the matrices or the maximum overall.

Cluster Shade

Local Homogeneity, Inverse Difference Moment (IDM) :

IDM is also influenced by the homogeneity of the image. Because of the weighting factor IDM will get small contributions from inhomogeneous areas. The result is a low IDM value for inhomogeneous images, and a relatively higher value for homogeneous images.

Sum of Squares, Variance :

This feature puts relatively high weights on the elements that differ from the average value of P(i, j).

Autocorrelation :

Other statistical approaches include an autocorrelation function, which has been used for analyzing the regularity. This function evaluates the linear spatial relationships between primitives. The set of autocorrelation coefficients shown below are used as texture features:

where p, q is the positional difference in the i, j direction, and M, N are image dimensions.

C.  Algorithm:-

Algorithm is design as per the following module that perform possible analysis for applied digital image of dermatitis skin patch by calculating above discussed features.

Load Image:

In this module we initially load the test image which  is to be classified. It can be done by using the uigetfile and imread functions in the matlab. To display the input image we use imshow function. After the loading the input image we will convert into gray scale image using rgb2gray function.

Preprocessing:

After the gray scale conversion we apply median filter to that image.

The Median filter is a nonlinear digital filtering technique, often used to remove noise. Such noise reduction is a typical pre-processing step to improve the results of later processing.  Median filtering is very widely used in digital image processing because, under certain conditions, it preserves edges while removing noise.

Feature Extraction:

The feature values are extracted from the filtered  test image .  We extract theGLCM feature for the input image. The Gray Level Co occurrence Matrix (GLCM) method is a way of extracting second order statistical texture features. A GLCM is a matrix where the number of rows and columns is equal to the number of gray levels in the image.

Texture Feature Extraction:

When the input data to an algorithm is too large to be processed and it is suspected to be notoriously redundant (much data, but not much information) then the input data will be transformed into a reduced representation set of features (also named features vector). Transforming the input data into the set of features is called feature extraction. The features provide the characteristics of the input type to the classifier by considering the description of the relevant properties of the image into a feature space. If the features extracted are carefully chosen, it is expected that they will extract the relevant information from the input data in order to perform the desired task using this reduced representation instead of the full size input.      Feature extraction involves simplifying the amount of resources required to describe a large set of data accurately. When performing  complex data one of the major problems stems from the number of variables involved.

GLCM

A gray level co-occurrence matrix (GLCM) contains information about the positions of pixels having similar gray level values.

A co-occurrence matrix is a two-dimensional array, P, in which both the rows and the columns represent a set of possible image values.

A GLCM Pd[i,j] is defined by first specifying a displacement vector d=(dx,dy) and counting all pairs of pixels separated by d having gray levels i and j.

The GLCM is defined by:

where nij is the number of occurrences of the pixel

values (i,j) lying at distance d in the image.

 The co-occurrence matrix Pd has dimension n× n,

where n is the number of gray levels in the image.

III. Result and Discussion

The developed advanced image segmentation algorithm for analysis and detection of  Dermatitis disease is useful for the classification of dermatitis disease using k-Means clustering method of segmentation  and can classify disease into 4 category based on the database stored in nntool and using fuzzy logic it matched the parameter of test image to the database image. This algorithm extract the 12 feature 12 database and a test image of patch of disease and compare them. On the basis best matched 5 feature it classify to one of the four category. In this paper analysis of Leprosy  test image is given. This algorithm first remove the noise of image and then perform k-means clustering image segmentation to extract features like energy, contrast, homogeneity etc and match with 12 featured of stored  database of  all images. Best 5 matched feature help to analyze and detect disease. The result of  best selected 5 features of 56 database images and feature of test image is shown in table1

Figure2:-Original Leprosy Image , Cluster 1 and cluster
Figure 3 :-Extracted Features of Database Images

Table 3. :- Feature Selection from database Images

Figure 4.:-Selected Features of Database Images

Table 4:-Selected features to classify Test Image in to Disease

Figure 5:-Regression Maodel

IV. Conclusion

In our research work mostly focus was given on three dermatitis diseases such as Leprosy, Psoriasis and Viitiligo but all other remaining diseases are covered under ‘other’ category based on similar features of remaining diseases. The design is first check with Minitab 16 regression model bur the regression value does not satisfy the standard value due to non linearity of database of extracted features. So the extracted features database of above four category is used for training, testing and validation using Artificial neural network ANN using nntool and nnstart tool which use for non linearity database also.  Initially Artificial Neural network is design with 165 sample set where number of neuron layes were 10 which produced regression result to less than 55. As number of sample data is increase with number of neuron layers, the regression value approaches to standard value 1. We got the highest regression value is 0.9 for sample set of 307 with number of neuron layers used are15. The Neural network is train with data base of 70% and test and validate with 15% each.

REFERENCES

  1. Edge Detection Techniques for Image Segmentation – A Survey of Soft Computing Approaches N. Senthilkumaran1 and R. Rajesh2 School of Computer Science and Engineering, Bharathiar University, Coimbatore -641 046, India. International Journal of Recent Trends in Engineering, Vol. 1, No. 2, May 2009
  2. D. Kwon et al. “A Image Segmentation Method Based on Improved Watershed           Algorithm and Region Merging,” IEEE Trans Circuits and Syst. Video Technol., Vol. 17, pp. 517 – 529, May 2007
  3. Bertelli, L., Sumengen, B., Manjunath, B.S., Gibou, F.: A variational framework for multiregion pairwise-similarity-based image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence 30(8) (2008) 1400{1415
  4. Payel Ghosh and Melanie Mitchell, “Segmentation of medical images using a genetic algorithm,” in GECCO, Mike Cattolico, Ed. 2006, pp. 1171–1178, ACM.
  5. Yuan, X., Situ, N., Zouridakis, G.: A narrow band graph partitioning method for skin lesion segmentation. Pattern Recognition 42 (2009) 1017-1028
  6. B. Erkol, R.H. Moss, R.J. Stanley, W.V. Stoecker, and E. Hvatum, “Automatic lesion boundary detection in dermoscopy images using gradient vector flow snakes,” Skin Research and Technology, vol. 11, pp. 17–26, 2005.
  7. J. Reinhardt and W. Higgins, “Automatis generation of image segmentation processes,” IEEE Trans. Image Processing, vol. 3, pp. 791–795, Nov1994.
  8. Silvio M. Pereira, Marco A. C. Frade, Rangaraj M. Rangayyan, Fellow, IEEE and Paulo M. Azevedo-Marques, Member, IEEE  “Classification of Color Images of Dermatological Ulcers”  IEEE journal of biomedical and health informatics, vol. 17, no. 1,pp. 136-142,  January 2013
  9.  I. Maglogiannis, S. Pavlopoulos, and D. Koutsouris, “An integrated computer supported acquisition, handling, and characterization system for pigmented skin lesions in dermatological images,” IEEE Trans. Inf. Technol.Biomed., vol. 9, no. 1, pp. 86–98, Mar. 2005.
  10. H. Oduncu, A. Hoppe, M. Clark, R. Williams, and K. Harding, “Analysis of skin wound images using digital color image processing: A preliminary communication,” Int. J. Low Extreme Wounds, vol. 3, no. 3, pp. 151–156, 2004.
  11. A. Tarallo, A. Gonzaga, and M. Frade, “Artificial neural networks applied to the segmentation and classification of digital images of cutaneous ulcers,” in Proc. IEEE 7th Int. Conf. Bioinformat. Bioeng., 2007, pp. 1–1.
  12. E. Dorileo, M. Frade, A. Roselino, R. Rangayyan, and P. Azevedo- Marques, “Color image processing and content-based image retrieval techniquesfor the analysis of dermatological lesions,” in Proc. 30th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc.. Vancouver, BC, Canada, Aug. 2008, pp. 1230–1233.
  13. R. Rangayyan, B. Acha, and C. Serrano. (2011). Color Image Processing With Biomedical Applications, ser. Press monograph 206. SPIE. [Online]. Available: http://books.google.co.in/books?id=a9CzuAAACAAJ
  14. R. Gonzalez and R. Woods, Digital Image Processing, 3rd ed. Englewood Cliffs, NJ: Pearson Prentice Hall, 2008.

Related posts

Chatbot At District Court

admin

Blood and organ Managment system

admin

FUEL MATE

admin

Leave a Comment